 It's great to see Ironic both in production at large scale today and innovating for the future. So it's important to remember that open source isn't just a name that we slap on things to make magic happen, there are well-known best practices that can make open source projects more successful and make companies more effective at participating. So our next speaker has a background in physics and biology as well as a deep understanding of technology which gives him a unique perspective on large complex systems and human collaboration. Please welcome Stefan Bucher. Good morning. Hello to the Open Infrastructure Summit here in Denver, howdy or servus as we say in the Bavarian Alps. You know it feels very much like home here. If you look to the left you can see the Mount Zugspitze which is the tallest mountain in Germany. On the right side you can see the Maroon Bells, very nice landscape here in Colorado so you understand why it really feels like home. And I had the chance by the way to spend yesterday at Lily Lake, a great place outdoors. It was just a bit windy and snowy but apart from that really great. You know I could continue now to share my passion with you for the remaining nine minutes on my outdoor skills but let's rather focus on technology and let me share my passion for technology with you. You know two most important topics that I want to broadcast to you is innovation and how you foster innovation and whilst you focus on innovation do not forget day two operations. To more specific you know we engineers, we love the cool stuff, I will share a bit around artificial intelligence around bare metal and machine learning. I will share with you some of our experience on CICD and the experience that we are having with containers. So let me start with artificial intelligence. So what have we done? You can imagine given the passion for the mountains. We basically trained our AI with masterpieces from Picasso, Moni and Van Gogh and basically then let it repaint the maroon bells and that's what the results going to look like. If you want to find out how the master would repaint you, you are more than welcome to join us at booth B4 around the coffee corner where you can really literally experience how the master would have repainted you. In order to run those huge applications based on TensorFlow, for sure we need huge infrastructure that our cloud platform basically offers. Apart from the massive scale and course and I will share some numbers with you later, we also host hardware accelerated flavors on our platform such as we hunt with GPS or FGPAs which allows us basically to host also high-performance computing applications which our customers highly appreciate. It's not only around hardware, the cool stuff, the engineering stuff, it's also around the software, it's around CICD. And for sure here we contribute significantly to the Sule project, but we're not only contributing to the project, we are also applying it and currently transforming our tests as well, Sule based in order to ensure that the integrity of our code is given. From a service perspective on our platform, one of the most asked for services are for sure containers, Kubernetes containers. We deploy them literally within minutes across three availability zones. We have currently several hundreds of those containers on our platform. If there is a requirement for special security, we also offer Cata containers to isolate this critical workload allowing for that additional security level. But all that nice and fancy stuff on innovation, may it be the containers, may it be machine learning is useless on our open telecom public cloud platform if we cannot ensure user reliability. It's key for us, we are at the scale right now where user reliability is at most critical and that's why we basically focus so much on day two operations based on German engineering. Want to share a bit more about that one with you? We currently grow our platform by around 400 servers per week, which translates into 12% growth. This is huge, this is enormous, you can imagine. And we heavily rely on our promoters based monitoring, ensuring that everything runs fine. We're not only using our promoters monitoring for the reactive part, but also for the proactive part to monitor capacity and ensure we're not running out of that one. Otherwise, without using this sophisticated monitoring on promoters, we would not be able to basically accommodate for that massive, massive growth. The growth on storage is even more significant. If we compare year over year, our block storage basically increased by more than 200%, we're currently hosting 11 petabyte of block storage. The object storage growth is even more impressive. It literally increased by 23 times if you do a year over year comparison. And right now, we're operating 46 terabyte of object storage, and that's growing like hell going forward. This growth is not only possible if you have proper monitoring, but also if you have a lot of automation in place, allowing you to handle that large volumes on our platform. Without that automation, we would not be able to cope with massive growth that we currently see. What do these numbers show? They obviously demonstrate that open stack basically scales. We currently operate 240,000 cores on our platform, which is huge. We are increasing our consumption, we embed consumption literally by 77%, our network transfer rates have significantly increased. So scale is massive, and thanks to all of you who contribute here to the open stack community helped us managing that massive scale and allow us basically operating on such a scale, our open telecom platform, which is, by the way, one of the largest public cloud providers over there in Europe, and we are continuing to grow. At the end, let me come back to my passion for the mountains, no matter if you're going hiking, no matter if how steep and difficult and dangerous a trail is, once you're on the top of the mountain, the few from the top all these rewards. So please enjoy innovation, but do not forget day two operations. Thank you so much. It really is great to see that open infrastructure really is used by companies like Deutsche Telecom in combination, so it isn't enough to just look at cloud in isolation or containers in isolation because the users are combining them together. Another important area of open infrastructure that's still very early in its evolution is 5G, and one of the reasons this is so interesting is that 5G involves so much more softer than the earlier generations, and especially so much more open source software.